[ https://issues.jboss.org/browse/JBJCA-1390?page=com.atlassian.jira.plugin... ] Stephen Fikes updated JBJCA-1390: --------------------------------- Description: The timeout counter increment performed in the [array list version of the pool|https://github.com/ironjacamar/ironjacamar/blob/ironjacamar-1.4.15.F...] is omitted from the [linked dequeue based version|https://github.com/ironjacamar/ironjacamar/blob/ironjacamar-1.4.1...]. This results in BlockingFailureCount not tracking IJ000655 errors. was: The Byteman rule below [1] (and a test-connection from the web console for the ExampleDS) verifies the finding that org.jboss.jca.core.connectionmanager.pool.mcp.SemaphoreConcurrentLinkedDequeManagedConnectionPool is the default pool now. As of the latest released code base (EAP 7.2 CP1), the timeout counter increment performed in the array list version of the pool [2] is omitted from the linked dequeue based version [3]. This looks like a bug. This results in BlockingFailureCount not tracking IJ000655 errors. [1] Byteman rule showing which managed connection pool implementations are created RULE org.jboss.jca.core.connectionmanager.pool.mcp.ManagedConnectionPool INTERFACE org.jboss.jca.core.connectionmanager.pool.mcp.ManagedConnectionPool METHOD <init> AT EXIT IF true DO traceStack("[BMAN] "); ENDRULE [2] https://github.com/ironjacamar/ironjacamar/blob/ironjacamar-1.4.15.Final/... [3] https://github.com/ironjacamar/ironjacamar/blob/ironjacamar-1.4.15.Final/... -- This message was sent by Atlassian Jira (v7.12.1#712002)